Performance of operator controlled systems is limited by our ability to measure system and component subsystem performance in a reliable and sensitive manner. Without adequate performance measures, there is no way to produce and test system designs, plan and execute training systems, or evaluate operational systems. Methods of developing these performance measures can be characterized by the way in which performance criteria are obtained. One approach which can be used when all factors that limit performance are known and quantified is an analytical method. For example, if a problem requires that an aircraft climb to a specified altitude while conserving fuel during the climb, the criterion, i.e., minimization of fuel, could be precisely defined analytically. Frequently, however, problems cannot be solved analytically, but demonstrations of superior as well as less than superior performances are available. In these cases an empirical approach can be used. This paper describes an empirical method for analyzing simulator flight data to develop weightings that permit performance discrimination between two groups of student pilots (one group of students successfully passed the initial Army rotary wing training program at Ft. Rucker, Alabama. The other group of students did not pass that training course.). The paper provides a description of the job sample (flight training) tests used to collect the data, the method for synthesizing the performance measures, and the results from using the measures to score student pilots.
